Instructions to use data-archetype/moss_image_base_dmd2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use data-archetype/moss_image_base_dmd2 with PEFT:
Task type is invalid.
- Diffusers
How to use data-archetype/moss_image_base_dmd2 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("OpenMOSS-Team/MOSS-Image-Base", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("data-archetype/moss_image_base_dmd2") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
Commit ·
969b5a3
0
Parent(s):
MOSS-Image-Base 8-step DMD2 LoRA
Browse files- .gitattributes +36 -0
- LICENSE +202 -0
- README.md +163 -0
- TECH_REPORT.md +115 -0
- images/example_goose.png +3 -0
- loras/moss_image_base_dmd2_bf16.safetensors +3 -0
- transformer/adapter_config.json +20 -0
- transformer/adapter_model.safetensors +3 -0
- workflows/moss_image_base_dmd2_8step.json +516 -0
- workflows/moss_image_base_dmd2_8step_api.json +121 -0
.gitattributes
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
images/example_goose.png filter=lfs diff=lfs merge=lfs -text
|
LICENSE
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
Apache License
|
| 3 |
+
Version 2.0, January 2004
|
| 4 |
+
http://www.apache.org/licenses/
|
| 5 |
+
|
| 6 |
+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
| 7 |
+
|
| 8 |
+
1. Definitions.
|
| 9 |
+
|
| 10 |
+
"License" shall mean the terms and conditions for use, reproduction,
|
| 11 |
+
and distribution as defined by Sections 1 through 9 of this document.
|
| 12 |
+
|
| 13 |
+
"Licensor" shall mean the copyright owner or entity authorized by
|
| 14 |
+
the copyright owner that is granting the License.
|
| 15 |
+
|
| 16 |
+
"Legal Entity" shall mean the union of the acting entity and all
|
| 17 |
+
other entities that control, are controlled by, or are under common
|
| 18 |
+
control with that entity. For the purposes of this definition,
|
| 19 |
+
"control" means (i) the power, direct or indirect, to cause the
|
| 20 |
+
direction or management of such entity, whether by contract or
|
| 21 |
+
otherwise, or (ii) ownership of fifty percent (50%) or more of the
|
| 22 |
+
outstanding shares, or (iii) beneficial ownership of such entity.
|
| 23 |
+
|
| 24 |
+
"You" (or "Your") shall mean an individual or Legal Entity
|
| 25 |
+
exercising permissions granted by this License.
|
| 26 |
+
|
| 27 |
+
"Source" form shall mean the preferred form for making modifications,
|
| 28 |
+
including but not limited to software source code, documentation
|
| 29 |
+
source, and configuration files.
|
| 30 |
+
|
| 31 |
+
"Object" form shall mean any form resulting from mechanical
|
| 32 |
+
transformation or translation of a Source form, including but
|
| 33 |
+
not limited to compiled object code, generated documentation,
|
| 34 |
+
and conversions to other media types.
|
| 35 |
+
|
| 36 |
+
"Work" shall mean the work of authorship, whether in Source or
|
| 37 |
+
Object form, made available under the License, as indicated by a
|
| 38 |
+
copyright notice that is included in or attached to the work
|
| 39 |
+
(an example is provided in the Appendix below).
|
| 40 |
+
|
| 41 |
+
"Derivative Works" shall mean any work, whether in Source or Object
|
| 42 |
+
form, that is based on (or derived from) the Work and for which the
|
| 43 |
+
editorial revisions, annotations, elaborations, or other modifications
|
| 44 |
+
represent, as a whole, an original work of authorship. For the purposes
|
| 45 |
+
of this License, Derivative Works shall not include works that remain
|
| 46 |
+
separable from, or merely link (or bind by name) to the interfaces of,
|
| 47 |
+
the Work and Derivative Works thereof.
|
| 48 |
+
|
| 49 |
+
"Contribution" shall mean any work of authorship, including
|
| 50 |
+
the original version of the Work and any modifications or additions
|
| 51 |
+
to that Work or Derivative Works thereof, that is intentionally
|
| 52 |
+
submitted to Licensor for inclusion in the Work by the copyright owner
|
| 53 |
+
or by an individual or Legal Entity authorized to submit on behalf of
|
| 54 |
+
the copyright owner. For the purposes of this definition, "submitted"
|
| 55 |
+
means any form of electronic, verbal, or written communication sent
|
| 56 |
+
to the Licensor or its representatives, including but not limited to
|
| 57 |
+
communication on electronic mailing lists, source code control systems,
|
| 58 |
+
and issue tracking systems that are managed by, or on behalf of, the
|
| 59 |
+
Licensor for the purpose of discussing and improving the Work, but
|
| 60 |
+
excluding communication that is conspicuously marked or otherwise
|
| 61 |
+
designated in writing by the copyright owner as "Not a Contribution."
|
| 62 |
+
|
| 63 |
+
"Contributor" shall mean Licensor and any individual or Legal Entity
|
| 64 |
+
on behalf of whom a Contribution has been received by Licensor and
|
| 65 |
+
subsequently incorporated within the Work.
|
| 66 |
+
|
| 67 |
+
2. Grant of Copyright License. Subject to the terms and conditions of
|
| 68 |
+
this License, each Contributor hereby grants to You a perpetual,
|
| 69 |
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
| 70 |
+
copyright license to reproduce, prepare Derivative Works of,
|
| 71 |
+
publicly display, publicly perform, sublicense, and distribute the
|
| 72 |
+
Work and such Derivative Works in Source or Object form.
|
| 73 |
+
|
| 74 |
+
3. Grant of Patent License. Subject to the terms and conditions of
|
| 75 |
+
this License, each Contributor hereby grants to You a perpetual,
|
| 76 |
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
| 77 |
+
(except as stated in this section) patent license to make, have made,
|
| 78 |
+
use, offer to sell, sell, import, and otherwise transfer the Work,
|
| 79 |
+
where such license applies only to those patent claims licensable
|
| 80 |
+
by such Contributor that are necessarily infringed by their
|
| 81 |
+
Contribution(s) alone or by combination of their Contribution(s)
|
| 82 |
+
with the Work to which such Contribution(s) was submitted. If You
|
| 83 |
+
institute patent litigation against any entity (including a
|
| 84 |
+
cross-claim or counterclaim in a lawsuit) alleging that the Work
|
| 85 |
+
or a Contribution incorporated within the Work constitutes direct
|
| 86 |
+
or contributory patent infringement, then any patent licenses
|
| 87 |
+
granted to You under this License for that Work shall terminate
|
| 88 |
+
as of the date such litigation is filed.
|
| 89 |
+
|
| 90 |
+
4. Redistribution. You may reproduce and distribute copies of the
|
| 91 |
+
Work or Derivative Works thereof in any medium, with or without
|
| 92 |
+
modifications, and in Source or Object form, provided that You
|
| 93 |
+
meet the following conditions:
|
| 94 |
+
|
| 95 |
+
(a) You must give any other recipients of the Work or
|
| 96 |
+
Derivative Works a copy of this License; and
|
| 97 |
+
|
| 98 |
+
(b) You must cause any modified files to carry prominent notices
|
| 99 |
+
stating that You changed the files; and
|
| 100 |
+
|
| 101 |
+
(c) You must retain, in the Source form of any Derivative Works
|
| 102 |
+
that You distribute, all copyright, patent, trademark, and
|
| 103 |
+
attribution notices from the Source form of the Work,
|
| 104 |
+
excluding those notices that do not pertain to any part of
|
| 105 |
+
the Derivative Works; and
|
| 106 |
+
|
| 107 |
+
(d) If the Work includes a "NOTICE" text file as part of its
|
| 108 |
+
distribution, then any Derivative Works that You distribute must
|
| 109 |
+
include a readable copy of the attribution notices contained
|
| 110 |
+
within such NOTICE file, excluding those notices that do not
|
| 111 |
+
pertain to any part of the Derivative Works, in at least one
|
| 112 |
+
of the following places: within a NOTICE text file distributed
|
| 113 |
+
as part of the Derivative Works; within the Source form or
|
| 114 |
+
documentation, if provided along with the Derivative Works; or,
|
| 115 |
+
within a display generated by the Derivative Works, if and
|
| 116 |
+
wherever such third-party notices normally appear. The contents
|
| 117 |
+
of the NOTICE file are for informational purposes only and
|
| 118 |
+
do not modify the License. You may add Your own attribution
|
| 119 |
+
notices within Derivative Works that You distribute, alongside
|
| 120 |
+
or as an addendum to the NOTICE text from the Work, provided
|
| 121 |
+
that such additional attribution notices cannot be construed
|
| 122 |
+
as modifying the License.
|
| 123 |
+
|
| 124 |
+
You may add Your own copyright statement to Your modifications and
|
| 125 |
+
may provide additional or different license terms and conditions
|
| 126 |
+
for use, reproduction, or distribution of Your modifications, or
|
| 127 |
+
for any such Derivative Works as a whole, provided Your use,
|
| 128 |
+
reproduction, and distribution of the Work otherwise complies with
|
| 129 |
+
the conditions stated in this License.
|
| 130 |
+
|
| 131 |
+
5. Submission of Contributions. Unless You explicitly state otherwise,
|
| 132 |
+
any Contribution intentionally submitted for inclusion in the Work
|
| 133 |
+
by You to the Licensor shall be under the terms and conditions of
|
| 134 |
+
this License, without any additional terms or conditions.
|
| 135 |
+
Notwithstanding the above, nothing herein shall supersede or modify
|
| 136 |
+
the terms of any separate license agreement you may have executed
|
| 137 |
+
with Licensor regarding such Contributions.
|
| 138 |
+
|
| 139 |
+
6. Trademarks. This License does not grant permission to use the trade
|
| 140 |
+
names, trademarks, service marks, or product names of the Licensor,
|
| 141 |
+
except as required for reasonable and customary use in describing the
|
| 142 |
+
origin of the Work and reproducing the content of the NOTICE file.
|
| 143 |
+
|
| 144 |
+
7. Disclaimer of Warranty. Unless required by applicable law or
|
| 145 |
+
agreed to in writing, Licensor provides the Work (and each
|
| 146 |
+
Contributor provides its Contributions) on an "AS IS" BASIS,
|
| 147 |
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
| 148 |
+
implied, including, without limitation, any warranties or conditions
|
| 149 |
+
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
|
| 150 |
+
PARTICULAR PURPOSE. You are solely responsible for determining the
|
| 151 |
+
appropriateness of using or redistributing the Work and assume any
|
| 152 |
+
risks associated with Your exercise of permissions under this License.
|
| 153 |
+
|
| 154 |
+
8. Limitation of Liability. In no event and under no legal theory,
|
| 155 |
+
whether in tort (including negligence), contract, or otherwise,
|
| 156 |
+
unless required by applicable law (such as deliberate and grossly
|
| 157 |
+
negligent acts) or agreed to in writing, shall any Contributor be
|
| 158 |
+
liable to You for damages, including any direct, indirect, special,
|
| 159 |
+
incidental, or consequential damages of any character arising as a
|
| 160 |
+
result of this License or out of the use or inability to use the
|
| 161 |
+
Work (including but not limited to damages for loss of goodwill,
|
| 162 |
+
work stoppage, computer failure or malfunction, or any and all
|
| 163 |
+
other commercial damages or losses), even if such Contributor
|
| 164 |
+
has been advised of the possibility of such damages.
|
| 165 |
+
|
| 166 |
+
9. Accepting Warranty or Additional Liability. While redistributing
|
| 167 |
+
the Work or Derivative Works thereof, You may choose to offer,
|
| 168 |
+
and charge a fee for, acceptance of support, warranty, indemnity,
|
| 169 |
+
or other liability obligations and/or rights consistent with this
|
| 170 |
+
License. However, in accepting such obligations, You may act only
|
| 171 |
+
on Your own behalf and on Your sole responsibility, not on behalf
|
| 172 |
+
of any other Contributor, and only if You agree to indemnify,
|
| 173 |
+
defend, and hold each Contributor harmless for any liability
|
| 174 |
+
incurred by, or claims asserted against, such Contributor by reason
|
| 175 |
+
of your accepting any such warranty or additional liability.
|
| 176 |
+
|
| 177 |
+
END OF TERMS AND CONDITIONS
|
| 178 |
+
|
| 179 |
+
APPENDIX: How to apply the Apache License to your work.
|
| 180 |
+
|
| 181 |
+
To apply the Apache License to your work, attach the following
|
| 182 |
+
boilerplate notice, with the fields enclosed by brackets "[]"
|
| 183 |
+
replaced with your own identifying information. (Don't include
|
| 184 |
+
the brackets!) The text should be enclosed in the appropriate
|
| 185 |
+
comment syntax for the file format. We also recommend that a
|
| 186 |
+
file or class name and description of purpose be included on the
|
| 187 |
+
same "printed page" as the copyright notice for easier
|
| 188 |
+
identification within third-party archives.
|
| 189 |
+
|
| 190 |
+
Copyright [yyyy] [name of copyright owner]
|
| 191 |
+
|
| 192 |
+
Licensed under the Apache License, Version 2.0 (the "License");
|
| 193 |
+
you may not use this file except in compliance with the License.
|
| 194 |
+
You may obtain a copy of the License at
|
| 195 |
+
|
| 196 |
+
http://www.apache.org/licenses/LICENSE-2.0
|
| 197 |
+
|
| 198 |
+
Unless required by applicable law or agreed to in writing, software
|
| 199 |
+
distributed under the License is distributed on an "AS IS" BASIS,
|
| 200 |
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 201 |
+
See the License for the specific language governing permissions and
|
| 202 |
+
limitations under the License.
|
README.md
ADDED
|
@@ -0,0 +1,163 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: OpenMOSS-Team/MOSS-Image-Base
|
| 3 |
+
base_model_relation: adapter
|
| 4 |
+
language:
|
| 5 |
+
- en
|
| 6 |
+
- zh
|
| 7 |
+
license: apache-2.0
|
| 8 |
+
library_name: peft
|
| 9 |
+
pipeline_tag: text-to-image
|
| 10 |
+
tags:
|
| 11 |
+
- moss-image
|
| 12 |
+
- diffusers
|
| 13 |
+
- lora
|
| 14 |
+
- text-to-image
|
| 15 |
+
- distillation
|
| 16 |
+
---
|
| 17 |
+
# moss_image_base_dmd2
|
| 18 |
+
|
| 19 |
+
An 8-step LoRA (rank 128) for [MOSS-Image-Base](https://huggingface.co/OpenMOSS-Team/MOSS-Image-Base), distilled by distribution
|
| 20 |
+
matching: 8 forwards at CFG 1, without a negative prompt, in place of the base model's
|
| 21 |
+
50 guided steps. It adapts only the transformer; the text encoder (Qwen3.5-9B), VAE
|
| 22 |
+
(FLUX.2 small decoder), tokenizer and inference pipeline are MOSS-Image-Base's,
|
| 23 |
+
unchanged. How it was trained is in [TECH_REPORT.md](TECH_REPORT.md).
|
| 24 |
+
|
| 25 |
+
It was trained on MOSS-Image-Base, but should work with most fine-tunes of it and with
|
| 26 |
+
other LoRAs: load it alongside them, at strength 1.
|
| 27 |
+
|
| 28 |
+

|
| 29 |
+
|
| 30 |
+
*8 steps at CFG 1, 1152 × 896, seed 42, with the prompt in the Diffusers example below. The image carries its ComfyUI workflow: drag it into ComfyUI to load it.*
|
| 31 |
+
|
| 32 |
+
| File | Format |
|
| 33 |
+
| --- | --- |
|
| 34 |
+
| `loras/moss_image_base_dmd2_bf16.safetensors` | ComfyUI LoRA (kohya keys on ComfyUI's module names, q/k/v as one LoRA on the fused `qkv`) |
|
| 35 |
+
| `transformer/` | PEFT LoRA adapter for the Diffusers `MOSSImageTransformer` |
|
| 36 |
+
| `images/example_goose.png` | Example image, its ComfyUI workflow embedded |
|
| 37 |
+
| `workflows/moss_image_base_dmd2_8step.json` | ComfyUI workflow (drag into ComfyUI) |
|
| 38 |
+
| `workflows/moss_image_base_dmd2_8step_api.json` | The same workflow in ComfyUI's API format |
|
| 39 |
+
|
| 40 |
+
Both LoRA files make the same weight change; the ComfyUI file packs each block's q, k and v
|
| 41 |
+
LoRAs into one exactly equivalent LoRA on ComfyUI's fused `qkv` projection.
|
| 42 |
+
|
| 43 |
+
## Sampling
|
| 44 |
+
|
| 45 |
+
The LoRA is trained for one sampler; others give worse results.
|
| 46 |
+
|
| 47 |
+
| Setting | Value |
|
| 48 |
+
| --- | --- |
|
| 49 |
+
| Steps | 8 |
|
| 50 |
+
| Sampler | At each step, take the model's clean estimate, then re-noise it with fresh noise to the next level (ComfyUI's `lcm`) |
|
| 51 |
+
| Schedule | 8 evenly spaced times 1, 7/8, ..., 1/8, then 0, under the flow shift 4.4817 (ComfyUI's `simple` scheduler after a ModelSamplingAuraFlow node at shift 4.4817); about 1, .969, .931, .882, .818, .729, .599, .39 |
|
| 52 |
+
| CFG | 1, no negative prompt (guidance is distilled in) |
|
| 53 |
+
| LoRA strength | 1 |
|
| 54 |
+
| Resolution | About 1 megapixel, any aspect ratio the base supports |
|
| 55 |
+
|
| 56 |
+
## ComfyUI
|
| 57 |
+
|
| 58 |
+
Use ComfyUI's native MOSS-Image support and load the workflow: drag `images/example_goose.png` or
|
| 59 |
+
`workflows/moss_image_base_dmd2_8step.json` into ComfyUI.
|
| 60 |
+
|
| 61 |
+
| File | Put in | Loader |
|
| 62 |
+
| --- | --- | --- |
|
| 63 |
+
| `loras/moss_image_base_dmd2_bf16.safetensors` (this repo) | `models/loras/` | LoraLoaderBypassModelOnly |
|
| 64 |
+
| `diffusion_models/moss_image_base_bf16.safetensors`, from Comfy-Org/MOSS-Image | `models/diffusion_models/` | Load Diffusion Model |
|
| 65 |
+
| `text_encoders/qwen3.5_9b_bf16.safetensors` (or its int8 / w4a8 variants), from Comfy-Org/MOSS-Image | `models/text_encoders/` | Load CLIP, type `moss_image` |
|
| 66 |
+
| `vae/flux2-vae_small.safetensors`, from Comfy-Org/MOSS-Image | `models/vae/` | Load VAE |
|
| 67 |
+
|
| 68 |
+
The workflow sets the sampler (`lcm`), scheduler (`simple`), 8 steps, CFG 1 and the
|
| 69 |
+
shift. ComfyUI's default MOSS-Image shift is 5, close to but not the 4.4817 the LoRA
|
| 70 |
+
was trained with; the workflow's ModelSamplingAuraFlow node sets it. Use that node (or
|
| 71 |
+
ModelSamplingFlux) to change the shift, not ModelSamplingSD3: it scales MOSS-Image's
|
| 72 |
+
timesteps differently and the output turns to noise.
|
| 73 |
+
|
| 74 |
+
Load the LoRA with the bypass loader ("Load LoRA (Bypass, Model Only)"), as the workflow does,
|
| 75 |
+
not the standard LoRA loader. The standard loader merges the LoRA into the bf16 weights, and
|
| 76 |
+
the weight change is smaller than bf16's rounding step, so most of it is lost. The bypass
|
| 77 |
+
loader applies it unmerged, as Diffusers does below.
|
| 78 |
+
|
| 79 |
+
## Diffusers
|
| 80 |
+
|
| 81 |
+
Install a PyTorch build that matches your CUDA driver first
|
| 82 |
+
([pytorch.org](https://pytorch.org/get-started/locally/)), then:
|
| 83 |
+
|
| 84 |
+
```bash
|
| 85 |
+
pip install "diffusers>=0.38" "transformers>=5.3" "accelerate>=1.10" "safetensors>=0.6" "peft>=0.21" "Pillow>=10"
|
| 86 |
+
```
|
| 87 |
+
|
| 88 |
+
```python
|
| 89 |
+
import math
|
| 90 |
+
|
| 91 |
+
import torch
|
| 92 |
+
from diffusers import AutoencoderKLFlux2, FlowMatchEulerDiscreteScheduler
|
| 93 |
+
from diffusers.utils.dynamic_modules_utils import get_class_from_dynamic_module
|
| 94 |
+
from peft import PeftModel
|
| 95 |
+
from transformers import AutoModel, AutoTokenizer
|
| 96 |
+
|
| 97 |
+
BASE = "OpenMOSS-Team/MOSS-Image-Base"
|
| 98 |
+
SHIFT = math.exp(1.5)
|
| 99 |
+
|
| 100 |
+
|
| 101 |
+
class FreshNoiseScheduler(FlowMatchEulerDiscreteScheduler):
|
| 102 |
+
"""The LoRA's sampler: jump to the clean estimate, re-noise with fresh noise."""
|
| 103 |
+
|
| 104 |
+
def set_timesteps(self, num_inference_steps=8, device=None):
|
| 105 |
+
t = torch.linspace(1.0, 1.0 / num_inference_steps, num_inference_steps)
|
| 106 |
+
sigmas = SHIFT * t / (1 + (SHIFT - 1) * t)
|
| 107 |
+
super().set_timesteps(device=device, sigmas=sigmas.tolist())
|
| 108 |
+
|
| 109 |
+
def step(self, model_output, timestep, sample, return_dict=True, **kwargs):
|
| 110 |
+
if self.step_index is None:
|
| 111 |
+
self._init_step_index(timestep)
|
| 112 |
+
sigma, sigma_next = self.sigmas[self.step_index], self.sigmas[self.step_index + 1]
|
| 113 |
+
clean = sample.float() - sigma * model_output.float()
|
| 114 |
+
if sigma_next > 0:
|
| 115 |
+
clean = (1 - sigma_next) * clean + sigma_next * torch.randn_like(clean)
|
| 116 |
+
self._step_index += 1
|
| 117 |
+
return (clean.to(sample.dtype),)
|
| 118 |
+
|
| 119 |
+
|
| 120 |
+
def moss_class(name):
|
| 121 |
+
return get_class_from_dynamic_module(
|
| 122 |
+
BASE, "pipeline_moss_image.py", class_name=name, trust_remote_code=True
|
| 123 |
+
)
|
| 124 |
+
|
| 125 |
+
|
| 126 |
+
transformer = moss_class("MOSSImageTransformer").from_pretrained(
|
| 127 |
+
"OpenMOSS-Team/MOSS-Image-Base", subfolder="transformer", torch_dtype=torch.bfloat16
|
| 128 |
+
)
|
| 129 |
+
transformer = PeftModel.from_pretrained(
|
| 130 |
+
transformer, "data-archetype/moss_image_base_dmd2", subfolder="transformer"
|
| 131 |
+
).get_base_model()
|
| 132 |
+
|
| 133 |
+
pipe = moss_class("MOSSImagePipeline")(
|
| 134 |
+
transformer=transformer,
|
| 135 |
+
scheduler=FreshNoiseScheduler(shift=1.0),
|
| 136 |
+
text_encoder=AutoModel.from_pretrained(BASE, subfolder="text_encoder", dtype=torch.bfloat16),
|
| 137 |
+
tokenizer=AutoTokenizer.from_pretrained(BASE, subfolder="tokenizer"),
|
| 138 |
+
vae=AutoencoderKLFlux2.from_pretrained(BASE, subfolder="vae", torch_dtype=torch.float32),
|
| 139 |
+
)
|
| 140 |
+
pipe.to("cuda") # or pipe.enable_model_cpu_offload() on a smaller GPU
|
| 141 |
+
|
| 142 |
+
torch.manual_seed(42)
|
| 143 |
+
image = pipe(
|
| 144 |
+
"A large brown and white goose resting on dry grass at the edge of a rocky pond, its patterned wing feathers folded neatly and its long neck curved as it looks toward the water. The sun casts warm light across the bird and the rough stone bank behind it.",
|
| 145 |
+
height=896,
|
| 146 |
+
width=1152,
|
| 147 |
+
num_inference_steps=8,
|
| 148 |
+
guidance_scale=1.0,
|
| 149 |
+
uncond_skip_layers=None,
|
| 150 |
+
).images[0]
|
| 151 |
+
```
|
| 152 |
+
|
| 153 |
+
The adapter is applied unmerged, at strength 1. Prompt formats (natural language, tags
|
| 154 |
+
and JSON layouts) and requirements are the base model's; see MOSS-Image-Base's
|
| 155 |
+
[model card](https://huggingface.co/OpenMOSS-Team/MOSS-Image-Base) and
|
| 156 |
+
[user guide](https://github.com/OpenMOSS/MOSS-Image/blob/main/USERGUIDE.md).
|
| 157 |
+
|
| 158 |
+
## License
|
| 159 |
+
|
| 160 |
+
The weights are released under the base model's license, Apache 2.0
|
| 161 |
+
([LICENSE](LICENSE), copied from OpenMOSS-Team/MOSS-Image-Base).
|
| 162 |
+
|
| 163 |
+
We thank the OpenMOSS team for MOSS-Image.
|
TECH_REPORT.md
ADDED
|
@@ -0,0 +1,115 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# moss_image_base_dmd2: technical report
|
| 2 |
+
|
| 3 |
+
An 8-step student of MOSS-Image-Base, trained by distribution matching in the style of
|
| 4 |
+
DMD2 (Yin et al., *Improved Distribution Matching Distillation for Fast Image
|
| 5 |
+
Synthesis*, 2024), without its GAN term. The student is a LoRA on the frozen
|
| 6 |
+
transformer; a second LoRA on the same frozen transformer learns the student's own
|
| 7 |
+
output distribution. No image enters a loss: training images only supply caption
|
| 8 |
+
texts and aspect ratios.
|
| 9 |
+
|
| 10 |
+
## Notation
|
| 11 |
+
|
| 12 |
+
Flow convention: `x_t = (1 - t) x0 + t eps`, with `t = 1` pure noise, `eps` standard
|
| 13 |
+
Gaussian noise.
|
| 14 |
+
|
| 15 |
+
| Symbol | Meaning |
|
| 16 |
+
| --- | --- |
|
| 17 |
+
| `S(x, t)` | Student: the frozen transformer plus the student LoRA, a velocity |
|
| 18 |
+
| `f(x, τ)` | Fake score: the frozen transformer plus the fake LoRA, a velocity |
|
| 19 |
+
| `u(x, τ)` | The frozen teacher on the prompt |
|
| 20 |
+
| `n(x, τ)` | The frozen teacher on the empty prompt with block 22 skipped |
|
| 21 |
+
| `v_real` | The guided teacher, `n + 3 (u - n)` |
|
| 22 |
+
| `x̂` | A clean estimate from the student |
|
| 23 |
+
| `τ` | The noise level at which teacher and fake are compared |
|
| 24 |
+
| `t_r` | Time of sampler step `r`, for `r` in 0..7 |
|
| 25 |
+
|
| 26 |
+
## Sampler
|
| 27 |
+
|
| 28 |
+
```
|
| 29 |
+
x = eps # t_0 = 1
|
| 30 |
+
for r in 0..7:
|
| 31 |
+
x̂ = x - t_r S(x, t_r)
|
| 32 |
+
if r < 7:
|
| 33 |
+
x = (1 - t_{r+1}) x̂ + t_{r+1} eps_r # fresh eps_r every step
|
| 34 |
+
output x̂
|
| 35 |
+
```
|
| 36 |
+
|
| 37 |
+
The times are 8 evenly spaced points 1, 7/8, ..., 1/8 under a log-SNR shift of -3,
|
| 38 |
+
`t_r = s u / (1 + (s - 1) u)` with `u = 1 - r/8` and `s = e^1.5 ≈ 4.4817`: about 1,
|
| 39 |
+
.969, .931, .882, .818, .729, .599, .39. Training uses exactly this chain, and
|
| 40 |
+
sampling should too.
|
| 41 |
+
|
| 42 |
+
## Objective
|
| 43 |
+
|
| 44 |
+
**Student update.** Run the chain without gradient up to step `r`, take step `r` with
|
| 45 |
+
gradient to get `x̂`, draw a noise level `τ`, noise `x_τ = (1 - τ) x̂ + τ eps`, and
|
| 46 |
+
evaluate `u`, `n` and `f` at `x_τ` without gradient. The force on `x̂` and its loss are
|
| 47 |
+
|
| 48 |
+
```
|
| 49 |
+
g = τ (v_real - f(x_τ, τ)) / N
|
| 50 |
+
N = mean | x̂ - (x_τ - τ v_real) | # per image, floored at 1e-6
|
| 51 |
+
L = 0.5 mean (x̂ - stop_grad(x̂ - g))²
|
| 52 |
+
```
|
| 53 |
+
|
| 54 |
+
so the gradient of `L` with respect to `x̂` is `g` (up to the element count): it moves
|
| 55 |
+
the student's output towards the guided teacher's clean estimate and away from the
|
| 56 |
+
fake's. `N`, DMD2's normaliser, reads the same guided teacher as the force.
|
| 57 |
+
|
| 58 |
+
**Fake update.** Run the student's chain without gradient to the output `x̂` of step
|
| 59 |
+
`r`, draw `τ`, noise with fresh `eps`, and train the fake LoRA by flow matching:
|
| 60 |
+
`mean (f(x_τ, τ) - (eps - x̂))²`. The fake sees only the noise level, not the step.
|
| 61 |
+
|
| 62 |
+
**Schedule of updates.** 200 fake-only updates first, then cycles of 5 fake updates
|
| 63 |
+
and 1 student update. A cycle trains one step `r`, shared by its fake and student
|
| 64 |
+
updates, and the next cycle takes the next step, 0 to 7 and round again. So the fake
|
| 65 |
+
has just trained on the outputs of the step it then judges.
|
| 66 |
+
|
| 67 |
+
**Noise levels.** Each update draws one `τ` per image, stratified across the update's
|
| 68 |
+
images: with `B` images, `q_i = (π(i) + ξ_i) / B` for a random permutation `π` and
|
| 69 |
+
uniform `ξ_i`, so each slice of the distribution gets exactly one image.
|
| 70 |
+
|
| 71 |
+
The student's `τ` is uniform and the fake's Beta(2, 2). Both are then shifted by a
|
| 72 |
+
log-SNR shift of -2.2 (`τ' = sigmoid(logit τ + 1.1)`) and not clamped.
|
| 73 |
+
|
| 74 |
+
Under that shift, the fake's Beta(2, 2) has a logit mean of about 1.10 and a standard
|
| 75 |
+
deviation of about 1.14. This is why it was chosen: it sits close to the noise levels
|
| 76 |
+
MOSS-Image itself was pretrained on, a truncated logit-normal with, in this
|
| 77 |
+
convention, a logit mean of 1.09 and a standard deviation of 0.95
|
| 78 |
+
([MOSS-Image technical report](https://github.com/OpenMOSS/MOSS-Image/blob/main/paper/MOSS_Image_Technical_Report.pdf)).
|
| 79 |
+
The distillation does not use that logit-normal itself.
|
| 80 |
+
|
| 81 |
+
## Recipe
|
| 82 |
+
|
| 83 |
+
| Setting | Value |
|
| 84 |
+
| --- | --- |
|
| 85 |
+
| Teacher | MOSS-Image-Base, frozen (bf16); text encoder and VAE frozen |
|
| 86 |
+
| Real score | CFG 3, the negative branch being the empty prompt with block 22 skipped, at every noise level |
|
| 87 |
+
| Student and fake | Two LoRAs on the same frozen transformer, each rank 128, alpha 128 (scale 1), A Kaiming-uniform and B zero at initialization, no dropout |
|
| 88 |
+
| Adapted modules | `q`, `k`, `v`, `o`, `ffn_up_gate` and `ffn_down` of all 42 transformer blocks, the time MLP's two linears, the shared block modulation and the head modulation; `ffn_up_gate` carries one rank-128 pair for each of its up and gate halves; the text embedder is not adapted |
|
| 89 |
+
| Optimizer | AdamW for each LoRA, β = (0.9, 0.95), ε 1e-8, no weight decay |
|
| 90 |
+
| Updates | 4,200: 200 fake-only, then 666 cycles of 5 fake and 1 student update (3,534 fake and 666 student updates in all) |
|
| 91 |
+
| Learning rate | 1e-5 for both LoRAs, no warmup: constant for the first 3,300 updates (516 student updates), then a 1 − √ decay to 0 over the last 900 (150 student updates). Both LoRAs follow the same decay, counted in updates of either kind |
|
| 92 |
+
| Gradient clipping | Global norm 10 |
|
| 93 |
+
| Precision | bf16 autocast for the forward and backward passes, TF32 for float32 matrix products; the LoRA factors, AdamW states and EMAs are float32 (the factors are cast to bf16 in the matrix product) |
|
| 94 |
+
| EMAs | Two power-function EMAs of the student's factors (Karras et al.), counted in student updates only: γ 6.94 with its decay capped at 0.9999, and γ 16.97 uncapped; the fake's factors have none |
|
| 95 |
+
| Batch | 16 images per update |
|
| 96 |
+
| Resolution | 29 aspect-ratio buckets of about 1 megapixel (1024 × 1024 and 28 others), sides in multiples of 64, aspect ratios from about 1 : 2.1 to 3.75 : 1 |
|
| 97 |
+
| Text | Captions 80% natural language, 20% JSON layouts (MOSS-Image's layout prompt format); prompts up to 2,048 tokens; no caption dropout |
|
| 98 |
+
| Released weights | an EMA of the student's factors (Karras power function, γ 6.94), stored in bfloat16 |
|
| 99 |
+
|
| 100 |
+
## Design notes
|
| 101 |
+
|
| 102 |
+
Every parameter and setting above was extensively ablated beforehand on an internal
|
| 103 |
+
2B text-to-image model with a similar architecture, at lower resolution. These are
|
| 104 |
+
the choices that mattered most, judged by eye on fixed prompts:
|
| 105 |
+
|
| 106 |
+
- **The step order matters most.** Giving the fake and the student independent
|
| 107 |
+
shuffles of the 8 steps produced wrong limbs, duplicated heads and objects within
|
| 108 |
+
a few hundred updates; sharing one step per cycle did not.
|
| 109 |
+
- **Equal learning rates.** A student slower than its fake delayed the early contrast
|
| 110 |
+
overshoot without lowering it.
|
| 111 |
+
- **Time and modulation layers in the LoRAs.** Adapting the time MLP and the
|
| 112 |
+
modulations brought the output close to the teacher's contrast much sooner.
|
| 113 |
+
- **Stratified noise levels** gave a lower contrast peak than independent draws.
|
| 114 |
+
- **Beta(2, 2) for the fake** gave cleaner structure than a uniform draw.
|
| 115 |
+
- **No GAN and no regression loss**: the distribution-matching force alone.
|
images/example_goose.png
ADDED
|
Git LFS Details
|
loras/moss_image_base_dmd2_bf16.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:36714b3fa11ad665cfc52f803fc6bed6ffedcd829700acc5938b0606957b4e5b
|
| 3 |
+
size 1012265104
|
transformer/adapter_config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model_name_or_path": "OpenMOSS-Team/MOSS-Image-Base",
|
| 3 |
+
"bias": "none",
|
| 4 |
+
"inference_mode": true,
|
| 5 |
+
"init_lora_weights": true,
|
| 6 |
+
"lora_alpha": 128,
|
| 7 |
+
"lora_dropout": 0.0,
|
| 8 |
+
"peft_type": "LORA",
|
| 9 |
+
"r": 128,
|
| 10 |
+
"rank_pattern": {
|
| 11 |
+
"ffn_up_gate": 256
|
| 12 |
+
},
|
| 13 |
+
"alpha_pattern": {
|
| 14 |
+
"ffn_up_gate": 256
|
| 15 |
+
},
|
| 16 |
+
"target_modules": "(layers\\.\\d+\\.proj\\.(q|k|v|o|ffn_up_gate|ffn_down)|time_embedding\\.(0|2)|shared_block_modulation|video_head_modulation)",
|
| 17 |
+
"task_type": null,
|
| 18 |
+
"use_dora": false,
|
| 19 |
+
"use_rslora": false
|
| 20 |
+
}
|
transformer/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e717a7aa558e8819b50c09e91efa49ba38ef2a44d5747477695cbb4cd72e786a
|
| 3 |
+
size 814086912
|
workflows/moss_image_base_dmd2_8step.json
ADDED
|
@@ -0,0 +1,516 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"last_node_id": 11,
|
| 3 |
+
"last_link_id": 11,
|
| 4 |
+
"nodes": [
|
| 5 |
+
{
|
| 6 |
+
"id": 1,
|
| 7 |
+
"type": "UNETLoader",
|
| 8 |
+
"pos": [
|
| 9 |
+
0,
|
| 10 |
+
0
|
| 11 |
+
],
|
| 12 |
+
"size": [
|
| 13 |
+
320,
|
| 14 |
+
110
|
| 15 |
+
],
|
| 16 |
+
"flags": {},
|
| 17 |
+
"order": 1,
|
| 18 |
+
"mode": 0,
|
| 19 |
+
"inputs": [],
|
| 20 |
+
"outputs": [
|
| 21 |
+
{
|
| 22 |
+
"name": "MODEL",
|
| 23 |
+
"type": "MODEL",
|
| 24 |
+
"links": [
|
| 25 |
+
1
|
| 26 |
+
],
|
| 27 |
+
"slot_index": 0
|
| 28 |
+
}
|
| 29 |
+
],
|
| 30 |
+
"properties": {
|
| 31 |
+
"Node name for S&R": "UNETLoader"
|
| 32 |
+
},
|
| 33 |
+
"widgets_values": [
|
| 34 |
+
"moss_image_base_bf16.safetensors",
|
| 35 |
+
"default"
|
| 36 |
+
]
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"id": 2,
|
| 40 |
+
"type": "LoraLoaderBypassModelOnly",
|
| 41 |
+
"pos": [
|
| 42 |
+
360,
|
| 43 |
+
0
|
| 44 |
+
],
|
| 45 |
+
"size": [
|
| 46 |
+
320,
|
| 47 |
+
110
|
| 48 |
+
],
|
| 49 |
+
"flags": {},
|
| 50 |
+
"order": 2,
|
| 51 |
+
"mode": 0,
|
| 52 |
+
"inputs": [
|
| 53 |
+
{
|
| 54 |
+
"name": "model",
|
| 55 |
+
"type": "MODEL",
|
| 56 |
+
"link": 1
|
| 57 |
+
}
|
| 58 |
+
],
|
| 59 |
+
"outputs": [
|
| 60 |
+
{
|
| 61 |
+
"name": "MODEL",
|
| 62 |
+
"type": "MODEL",
|
| 63 |
+
"links": [
|
| 64 |
+
2
|
| 65 |
+
],
|
| 66 |
+
"slot_index": 0
|
| 67 |
+
}
|
| 68 |
+
],
|
| 69 |
+
"properties": {
|
| 70 |
+
"Node name for S&R": "LoraLoaderBypassModelOnly"
|
| 71 |
+
},
|
| 72 |
+
"widgets_values": [
|
| 73 |
+
"moss_image_base_dmd2_bf16.safetensors",
|
| 74 |
+
1.0
|
| 75 |
+
]
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"id": 3,
|
| 79 |
+
"type": "ModelSamplingAuraFlow",
|
| 80 |
+
"pos": [
|
| 81 |
+
720,
|
| 82 |
+
0
|
| 83 |
+
],
|
| 84 |
+
"size": [
|
| 85 |
+
320,
|
| 86 |
+
110
|
| 87 |
+
],
|
| 88 |
+
"flags": {},
|
| 89 |
+
"order": 3,
|
| 90 |
+
"mode": 0,
|
| 91 |
+
"inputs": [
|
| 92 |
+
{
|
| 93 |
+
"name": "model",
|
| 94 |
+
"type": "MODEL",
|
| 95 |
+
"link": 2
|
| 96 |
+
}
|
| 97 |
+
],
|
| 98 |
+
"outputs": [
|
| 99 |
+
{
|
| 100 |
+
"name": "MODEL",
|
| 101 |
+
"type": "MODEL",
|
| 102 |
+
"links": [
|
| 103 |
+
3
|
| 104 |
+
],
|
| 105 |
+
"slot_index": 0
|
| 106 |
+
}
|
| 107 |
+
],
|
| 108 |
+
"properties": {
|
| 109 |
+
"Node name for S&R": "ModelSamplingAuraFlow"
|
| 110 |
+
},
|
| 111 |
+
"widgets_values": [
|
| 112 |
+
4.4817
|
| 113 |
+
],
|
| 114 |
+
"title": "Flow shift e^1.5 (as trained)"
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"id": 4,
|
| 118 |
+
"type": "CLIPLoader",
|
| 119 |
+
"pos": [
|
| 120 |
+
0,
|
| 121 |
+
180
|
| 122 |
+
],
|
| 123 |
+
"size": [
|
| 124 |
+
320,
|
| 125 |
+
110
|
| 126 |
+
],
|
| 127 |
+
"flags": {},
|
| 128 |
+
"order": 4,
|
| 129 |
+
"mode": 0,
|
| 130 |
+
"inputs": [],
|
| 131 |
+
"outputs": [
|
| 132 |
+
{
|
| 133 |
+
"name": "CLIP",
|
| 134 |
+
"type": "CLIP",
|
| 135 |
+
"links": [
|
| 136 |
+
4
|
| 137 |
+
],
|
| 138 |
+
"slot_index": 0
|
| 139 |
+
}
|
| 140 |
+
],
|
| 141 |
+
"properties": {
|
| 142 |
+
"Node name for S&R": "CLIPLoader"
|
| 143 |
+
},
|
| 144 |
+
"widgets_values": [
|
| 145 |
+
"qwen3.5_9b_bf16.safetensors",
|
| 146 |
+
"moss_image",
|
| 147 |
+
"default"
|
| 148 |
+
]
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"id": 5,
|
| 152 |
+
"type": "VAELoader",
|
| 153 |
+
"pos": [
|
| 154 |
+
0,
|
| 155 |
+
360
|
| 156 |
+
],
|
| 157 |
+
"size": [
|
| 158 |
+
320,
|
| 159 |
+
110
|
| 160 |
+
],
|
| 161 |
+
"flags": {},
|
| 162 |
+
"order": 5,
|
| 163 |
+
"mode": 0,
|
| 164 |
+
"inputs": [],
|
| 165 |
+
"outputs": [
|
| 166 |
+
{
|
| 167 |
+
"name": "VAE",
|
| 168 |
+
"type": "VAE",
|
| 169 |
+
"links": [
|
| 170 |
+
10
|
| 171 |
+
],
|
| 172 |
+
"slot_index": 0
|
| 173 |
+
}
|
| 174 |
+
],
|
| 175 |
+
"properties": {
|
| 176 |
+
"Node name for S&R": "VAELoader"
|
| 177 |
+
},
|
| 178 |
+
"widgets_values": [
|
| 179 |
+
"flux2-vae_small.safetensors"
|
| 180 |
+
]
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"id": 6,
|
| 184 |
+
"type": "CLIPTextEncode",
|
| 185 |
+
"pos": [
|
| 186 |
+
360,
|
| 187 |
+
180
|
| 188 |
+
],
|
| 189 |
+
"size": [
|
| 190 |
+
320,
|
| 191 |
+
110
|
| 192 |
+
],
|
| 193 |
+
"flags": {},
|
| 194 |
+
"order": 6,
|
| 195 |
+
"mode": 0,
|
| 196 |
+
"inputs": [
|
| 197 |
+
{
|
| 198 |
+
"name": "clip",
|
| 199 |
+
"type": "CLIP",
|
| 200 |
+
"link": 4
|
| 201 |
+
}
|
| 202 |
+
],
|
| 203 |
+
"outputs": [
|
| 204 |
+
{
|
| 205 |
+
"name": "CONDITIONING",
|
| 206 |
+
"type": "CONDITIONING",
|
| 207 |
+
"links": [
|
| 208 |
+
5,
|
| 209 |
+
6
|
| 210 |
+
],
|
| 211 |
+
"slot_index": 0
|
| 212 |
+
}
|
| 213 |
+
],
|
| 214 |
+
"properties": {
|
| 215 |
+
"Node name for S&R": "CLIPTextEncode"
|
| 216 |
+
},
|
| 217 |
+
"widgets_values": [
|
| 218 |
+
"A large brown and white goose resting on dry grass at the edge of a rocky pond, its patterned wing feathers folded neatly and its long neck curved as it looks toward the water. The sun casts warm light across the bird and the rough stone bank behind it."
|
| 219 |
+
]
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"id": 7,
|
| 223 |
+
"type": "ConditioningZeroOut",
|
| 224 |
+
"pos": [
|
| 225 |
+
720,
|
| 226 |
+
180
|
| 227 |
+
],
|
| 228 |
+
"size": [
|
| 229 |
+
320,
|
| 230 |
+
110
|
| 231 |
+
],
|
| 232 |
+
"flags": {},
|
| 233 |
+
"order": 7,
|
| 234 |
+
"mode": 0,
|
| 235 |
+
"inputs": [
|
| 236 |
+
{
|
| 237 |
+
"name": "conditioning",
|
| 238 |
+
"type": "CONDITIONING",
|
| 239 |
+
"link": 6
|
| 240 |
+
}
|
| 241 |
+
],
|
| 242 |
+
"outputs": [
|
| 243 |
+
{
|
| 244 |
+
"name": "CONDITIONING",
|
| 245 |
+
"type": "CONDITIONING",
|
| 246 |
+
"links": [
|
| 247 |
+
7
|
| 248 |
+
],
|
| 249 |
+
"slot_index": 0
|
| 250 |
+
}
|
| 251 |
+
],
|
| 252 |
+
"properties": {
|
| 253 |
+
"Node name for S&R": "ConditioningZeroOut"
|
| 254 |
+
},
|
| 255 |
+
"widgets_values": [],
|
| 256 |
+
"title": "Unused at CFG 1"
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"id": 8,
|
| 260 |
+
"type": "EmptyFlux2LatentImage",
|
| 261 |
+
"pos": [
|
| 262 |
+
360,
|
| 263 |
+
360
|
| 264 |
+
],
|
| 265 |
+
"size": [
|
| 266 |
+
320,
|
| 267 |
+
110
|
| 268 |
+
],
|
| 269 |
+
"flags": {},
|
| 270 |
+
"order": 8,
|
| 271 |
+
"mode": 0,
|
| 272 |
+
"inputs": [],
|
| 273 |
+
"outputs": [
|
| 274 |
+
{
|
| 275 |
+
"name": "LATENT",
|
| 276 |
+
"type": "LATENT",
|
| 277 |
+
"links": [
|
| 278 |
+
8
|
| 279 |
+
],
|
| 280 |
+
"slot_index": 0
|
| 281 |
+
}
|
| 282 |
+
],
|
| 283 |
+
"properties": {
|
| 284 |
+
"Node name for S&R": "EmptyFlux2LatentImage"
|
| 285 |
+
},
|
| 286 |
+
"widgets_values": [
|
| 287 |
+
1152,
|
| 288 |
+
896,
|
| 289 |
+
1
|
| 290 |
+
]
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"id": 9,
|
| 294 |
+
"type": "KSampler",
|
| 295 |
+
"pos": [
|
| 296 |
+
1080,
|
| 297 |
+
0
|
| 298 |
+
],
|
| 299 |
+
"size": [
|
| 300 |
+
320,
|
| 301 |
+
110
|
| 302 |
+
],
|
| 303 |
+
"flags": {},
|
| 304 |
+
"order": 9,
|
| 305 |
+
"mode": 0,
|
| 306 |
+
"inputs": [
|
| 307 |
+
{
|
| 308 |
+
"name": "model",
|
| 309 |
+
"type": "MODEL",
|
| 310 |
+
"link": 3
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"name": "positive",
|
| 314 |
+
"type": "CONDITIONING",
|
| 315 |
+
"link": 5
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"name": "negative",
|
| 319 |
+
"type": "CONDITIONING",
|
| 320 |
+
"link": 7
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"name": "latent_image",
|
| 324 |
+
"type": "LATENT",
|
| 325 |
+
"link": 8
|
| 326 |
+
}
|
| 327 |
+
],
|
| 328 |
+
"outputs": [
|
| 329 |
+
{
|
| 330 |
+
"name": "LATENT",
|
| 331 |
+
"type": "LATENT",
|
| 332 |
+
"links": [
|
| 333 |
+
9
|
| 334 |
+
],
|
| 335 |
+
"slot_index": 0
|
| 336 |
+
}
|
| 337 |
+
],
|
| 338 |
+
"properties": {
|
| 339 |
+
"Node name for S&R": "KSampler"
|
| 340 |
+
},
|
| 341 |
+
"widgets_values": [
|
| 342 |
+
42,
|
| 343 |
+
"fixed",
|
| 344 |
+
8,
|
| 345 |
+
1.0,
|
| 346 |
+
"lcm",
|
| 347 |
+
"simple",
|
| 348 |
+
1.0
|
| 349 |
+
]
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"id": 10,
|
| 353 |
+
"type": "VAEDecode",
|
| 354 |
+
"pos": [
|
| 355 |
+
1440,
|
| 356 |
+
0
|
| 357 |
+
],
|
| 358 |
+
"size": [
|
| 359 |
+
320,
|
| 360 |
+
110
|
| 361 |
+
],
|
| 362 |
+
"flags": {},
|
| 363 |
+
"order": 10,
|
| 364 |
+
"mode": 0,
|
| 365 |
+
"inputs": [
|
| 366 |
+
{
|
| 367 |
+
"name": "samples",
|
| 368 |
+
"type": "LATENT",
|
| 369 |
+
"link": 9
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"name": "vae",
|
| 373 |
+
"type": "VAE",
|
| 374 |
+
"link": 10
|
| 375 |
+
}
|
| 376 |
+
],
|
| 377 |
+
"outputs": [
|
| 378 |
+
{
|
| 379 |
+
"name": "IMAGE",
|
| 380 |
+
"type": "IMAGE",
|
| 381 |
+
"links": [
|
| 382 |
+
11
|
| 383 |
+
],
|
| 384 |
+
"slot_index": 0
|
| 385 |
+
}
|
| 386 |
+
],
|
| 387 |
+
"properties": {
|
| 388 |
+
"Node name for S&R": "VAEDecode"
|
| 389 |
+
},
|
| 390 |
+
"widgets_values": []
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"id": 11,
|
| 394 |
+
"type": "SaveImage",
|
| 395 |
+
"pos": [
|
| 396 |
+
1440,
|
| 397 |
+
180
|
| 398 |
+
],
|
| 399 |
+
"size": [
|
| 400 |
+
320,
|
| 401 |
+
110
|
| 402 |
+
],
|
| 403 |
+
"flags": {},
|
| 404 |
+
"order": 11,
|
| 405 |
+
"mode": 0,
|
| 406 |
+
"inputs": [
|
| 407 |
+
{
|
| 408 |
+
"name": "images",
|
| 409 |
+
"type": "IMAGE",
|
| 410 |
+
"link": 11
|
| 411 |
+
}
|
| 412 |
+
],
|
| 413 |
+
"outputs": [],
|
| 414 |
+
"properties": {
|
| 415 |
+
"Node name for S&R": "SaveImage"
|
| 416 |
+
},
|
| 417 |
+
"widgets_values": [
|
| 418 |
+
"moss_image_base_dmd2"
|
| 419 |
+
]
|
| 420 |
+
}
|
| 421 |
+
],
|
| 422 |
+
"links": [
|
| 423 |
+
[
|
| 424 |
+
1,
|
| 425 |
+
1,
|
| 426 |
+
0,
|
| 427 |
+
2,
|
| 428 |
+
0,
|
| 429 |
+
"MODEL"
|
| 430 |
+
],
|
| 431 |
+
[
|
| 432 |
+
2,
|
| 433 |
+
2,
|
| 434 |
+
0,
|
| 435 |
+
3,
|
| 436 |
+
0,
|
| 437 |
+
"MODEL"
|
| 438 |
+
],
|
| 439 |
+
[
|
| 440 |
+
3,
|
| 441 |
+
3,
|
| 442 |
+
0,
|
| 443 |
+
9,
|
| 444 |
+
0,
|
| 445 |
+
"MODEL"
|
| 446 |
+
],
|
| 447 |
+
[
|
| 448 |
+
4,
|
| 449 |
+
4,
|
| 450 |
+
0,
|
| 451 |
+
6,
|
| 452 |
+
0,
|
| 453 |
+
"CLIP"
|
| 454 |
+
],
|
| 455 |
+
[
|
| 456 |
+
5,
|
| 457 |
+
6,
|
| 458 |
+
0,
|
| 459 |
+
9,
|
| 460 |
+
1,
|
| 461 |
+
"CONDITIONING"
|
| 462 |
+
],
|
| 463 |
+
[
|
| 464 |
+
6,
|
| 465 |
+
6,
|
| 466 |
+
0,
|
| 467 |
+
7,
|
| 468 |
+
0,
|
| 469 |
+
"CONDITIONING"
|
| 470 |
+
],
|
| 471 |
+
[
|
| 472 |
+
7,
|
| 473 |
+
7,
|
| 474 |
+
0,
|
| 475 |
+
9,
|
| 476 |
+
2,
|
| 477 |
+
"CONDITIONING"
|
| 478 |
+
],
|
| 479 |
+
[
|
| 480 |
+
8,
|
| 481 |
+
8,
|
| 482 |
+
0,
|
| 483 |
+
9,
|
| 484 |
+
3,
|
| 485 |
+
"LATENT"
|
| 486 |
+
],
|
| 487 |
+
[
|
| 488 |
+
9,
|
| 489 |
+
9,
|
| 490 |
+
0,
|
| 491 |
+
10,
|
| 492 |
+
0,
|
| 493 |
+
"LATENT"
|
| 494 |
+
],
|
| 495 |
+
[
|
| 496 |
+
10,
|
| 497 |
+
5,
|
| 498 |
+
0,
|
| 499 |
+
10,
|
| 500 |
+
1,
|
| 501 |
+
"VAE"
|
| 502 |
+
],
|
| 503 |
+
[
|
| 504 |
+
11,
|
| 505 |
+
10,
|
| 506 |
+
0,
|
| 507 |
+
11,
|
| 508 |
+
0,
|
| 509 |
+
"IMAGE"
|
| 510 |
+
]
|
| 511 |
+
],
|
| 512 |
+
"groups": [],
|
| 513 |
+
"config": {},
|
| 514 |
+
"extra": {},
|
| 515 |
+
"version": 0.4
|
| 516 |
+
}
|
workflows/moss_image_base_dmd2_8step_api.json
ADDED
|
@@ -0,0 +1,121 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"1": {
|
| 3 |
+
"class_type": "UNETLoader",
|
| 4 |
+
"inputs": {
|
| 5 |
+
"unet_name": "moss_image_base_bf16.safetensors",
|
| 6 |
+
"weight_dtype": "default"
|
| 7 |
+
}
|
| 8 |
+
},
|
| 9 |
+
"2": {
|
| 10 |
+
"class_type": "LoraLoaderBypassModelOnly",
|
| 11 |
+
"inputs": {
|
| 12 |
+
"model": [
|
| 13 |
+
"1",
|
| 14 |
+
0
|
| 15 |
+
],
|
| 16 |
+
"lora_name": "moss_image_base_dmd2_bf16.safetensors",
|
| 17 |
+
"strength_model": 1.0
|
| 18 |
+
}
|
| 19 |
+
},
|
| 20 |
+
"3": {
|
| 21 |
+
"class_type": "ModelSamplingAuraFlow",
|
| 22 |
+
"inputs": {
|
| 23 |
+
"model": [
|
| 24 |
+
"2",
|
| 25 |
+
0
|
| 26 |
+
],
|
| 27 |
+
"shift": 4.4817
|
| 28 |
+
}
|
| 29 |
+
},
|
| 30 |
+
"4": {
|
| 31 |
+
"class_type": "CLIPLoader",
|
| 32 |
+
"inputs": {
|
| 33 |
+
"clip_name": "qwen3.5_9b_bf16.safetensors",
|
| 34 |
+
"type": "moss_image",
|
| 35 |
+
"device": "default"
|
| 36 |
+
}
|
| 37 |
+
},
|
| 38 |
+
"5": {
|
| 39 |
+
"class_type": "VAELoader",
|
| 40 |
+
"inputs": {
|
| 41 |
+
"vae_name": "flux2-vae_small.safetensors"
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"6": {
|
| 45 |
+
"class_type": "CLIPTextEncode",
|
| 46 |
+
"inputs": {
|
| 47 |
+
"clip": [
|
| 48 |
+
"4",
|
| 49 |
+
0
|
| 50 |
+
],
|
| 51 |
+
"text": "A large brown and white goose resting on dry grass at the edge of a rocky pond, its patterned wing feathers folded neatly and its long neck curved as it looks toward the water. The sun casts warm light across the bird and the rough stone bank behind it."
|
| 52 |
+
}
|
| 53 |
+
},
|
| 54 |
+
"7": {
|
| 55 |
+
"class_type": "ConditioningZeroOut",
|
| 56 |
+
"inputs": {
|
| 57 |
+
"conditioning": [
|
| 58 |
+
"6",
|
| 59 |
+
0
|
| 60 |
+
]
|
| 61 |
+
}
|
| 62 |
+
},
|
| 63 |
+
"8": {
|
| 64 |
+
"class_type": "EmptyFlux2LatentImage",
|
| 65 |
+
"inputs": {
|
| 66 |
+
"width": 1152,
|
| 67 |
+
"height": 896,
|
| 68 |
+
"batch_size": 1
|
| 69 |
+
}
|
| 70 |
+
},
|
| 71 |
+
"9": {
|
| 72 |
+
"class_type": "KSampler",
|
| 73 |
+
"inputs": {
|
| 74 |
+
"model": [
|
| 75 |
+
"3",
|
| 76 |
+
0
|
| 77 |
+
],
|
| 78 |
+
"positive": [
|
| 79 |
+
"6",
|
| 80 |
+
0
|
| 81 |
+
],
|
| 82 |
+
"negative": [
|
| 83 |
+
"7",
|
| 84 |
+
0
|
| 85 |
+
],
|
| 86 |
+
"latent_image": [
|
| 87 |
+
"8",
|
| 88 |
+
0
|
| 89 |
+
],
|
| 90 |
+
"seed": 42,
|
| 91 |
+
"steps": 8,
|
| 92 |
+
"cfg": 1.0,
|
| 93 |
+
"sampler_name": "lcm",
|
| 94 |
+
"scheduler": "simple",
|
| 95 |
+
"denoise": 1.0
|
| 96 |
+
}
|
| 97 |
+
},
|
| 98 |
+
"10": {
|
| 99 |
+
"class_type": "VAEDecode",
|
| 100 |
+
"inputs": {
|
| 101 |
+
"samples": [
|
| 102 |
+
"9",
|
| 103 |
+
0
|
| 104 |
+
],
|
| 105 |
+
"vae": [
|
| 106 |
+
"5",
|
| 107 |
+
0
|
| 108 |
+
]
|
| 109 |
+
}
|
| 110 |
+
},
|
| 111 |
+
"11": {
|
| 112 |
+
"class_type": "SaveImage",
|
| 113 |
+
"inputs": {
|
| 114 |
+
"images": [
|
| 115 |
+
"10",
|
| 116 |
+
0
|
| 117 |
+
],
|
| 118 |
+
"filename_prefix": "moss_image_base_dmd2"
|
| 119 |
+
}
|
| 120 |
+
}
|
| 121 |
+
}
|